This thesis describes application of image processing for precise distance measurement in self acting production of a tip for AFM microscopes. The main goal is to measure distances between assembly parts during fabrication process. The purpose is to acquire a data for self acting assembly line which have to substitute inaccurate and nonrecurring manual assembly process. The assembly process consists of three technological steps. In first two steps the tungsten wire is glued to the cantilever. Distance measurement is necessary in all axes for proper alignment of parts. In third step the sharp tip is etched by KOH solution. The right distance between liquid level and the cantilever must be kept. A camera with high resolution and macro objective is used to acquire an image. Camera image is then calibrated to suppress distortions and scene position with respect to camera position. Length conversion coefficient is also computed. Object recognition and distance measurement is based on standard computer vision methods, mainly: adaptive thresholding, moments, image statistics, canny edge detector, Hough transform… Proposed algorithms have been implemented in C++ using Intel OpenCV library. The final achieved distance resolution is about 10µm per pixel. Algorithm output was successfully used to assembly few test tips.